Tuned Sound Amplifier/Receiver for the VLF Christmas Eve transmission from SAQ, Grimeton, Sweden

Jump to Latest
#144
Not to worry.

For the sake of completeness here are spectrum this morning the first when SAQ not transmitting the second when SAQ transmitting. The SAQ peak is very sharp just one frequency - very exact.

Spectrum SAQ not transmitting.png


Spectrum SAQ transmitting.png
 
#145
Congrats, well done! It's unlikely much can be done about the noise. It has the same sound signature as what I hear at this location, with the difference that it was much stronger due to the presence of thunderstorm activity, supposed to be absent in the dry season. So I heard and saw nothing but that noise, even at BW 10 Hz (sound) and 1 Hz (display).
1703416057585.png
 
#148
fowlay said:
Listening to a recording made with e.g. Audacity doesn't work for me - I am too old to hear the 17200 Hz tone.

Playing it back at, say, half speed should work in principle. However, there may be other VLF transmitters that drown out the SAQ message. These other transmitters may be as close as 1000 Hz away from 17200 Hz, so an antenna resonant at 17200 Hz will help. And bandpass filtering may improve the audibility further - an audio bandwidth of 200 Hz or less should be ideal.

Playing at reduced speed may also make it easier to work out the message. The telegraphist at SAQ sends at around 15 WPM (words per minute) - a bit fast for me; I can decipher it, but only by playing it over and over many times.
Click to expand...


Thanks for your help, playing the file back slow works.

After a few attempts this seemed to be the best way to process:-

1. Cntrl A to select all the file
2. Effect - Filter - Filter Curve EQ (draw a rectangular curve -50dB everywhere except between 17000 and 17500 Hz where +6dB was good).
3. Then check spectrum Go Analyse - Plot spectrum
4. Decimate any strong remaining non SAQ peaks by using notch filter Effect - EQ - Notch
5. To change the sound from 17200 Hz to a low frequency Go to Effect - Pitch and Tempo - Change Pitch
6. in frequency menu select "from 17200 to 516 Hz" then Apply.

The Morse code is now at 516 Hz.
 
#149
richgwilliams said:
Upon reflection I would prefer to use a better OpAmp. The 300K input resistor and the OpAmp I used contribute to noise. A better JFET type OpAmp as a non inverting amp with low gain would feel better. It would also be better to have more analogue bandpass filtering. The antenna I have is not big enough (high enough) really to do a good job.

TI do slate the LM358 now, probably because they have released an upgraded part, they have published a really detailed critique of the LM358! I often use the Analog Devices AD820, its a good general purpose FET amp with good noise figure and open loop gain of 40dB at 20 KHz.

Its hard to say what noise is coming from where at the moment,
Click to expand...
Congratulations for your good project. FYI, OP27 seems to be a favourite for boosting mystery signals. PC audio inputs are noisy so something like the Focusrite Scarlett Solo, with USB connection, gives a clean signal from the gear being tested.
Connecting a PC to the signal receiver will probably cause ground loops.

Here is a typical circuit for VLF. The opamps are biased for single supply operation so that a 12vdc battery can be used.
 

Attachments

  • VLF-dipole.png
    VLF-dipole.png
    18.1 KB · Views: 108
#155
fubar3 said:
FYI, OP27 seems to be a favourite for boosting mystery signals
Click to expand...
Yes the OP27 is loosing some of its open loop gain and common mode rejection at 20kHz, also it is really for +- 15 volt supplies (there is the OP37, faster but still +- 15v). The LM385, TL084 and AD820 are all loosing some open loop gain and common mode rejection at 20kHz. There are fet/jfet OpAmps that are faster and better suited.

For a VLF receiver looks like a jfet or even a small mosfet makes a good front end. Also a good RF bipolar transistor connected as emitter follower would work well, one of the old Germanium RF transistors (eg AF116) would also work, then there are special purpose IC's

The SAQ carrier seems held really close to 17200 Hz and appears as a very sharp peak in the spectrum, good analog sharp bandpass filters with a very narrow pass band centred on 17200 Hz will help to receive this signal anywhere globally, no mixing just record at 48k Samples per second then let digital processing do the rest.
 
Last edited:
#159
Talking about things forgotten I remember now that transistor radio medium wave Ferrite Rods had the main resonant winding and also a second winding of much fewer turns that fed into the first RF transistor. Similarly the Eddystone circuit mentioned above had a tuned input circuit (connected to Antenna) with a tapping of only a few turns feeding into the emitter of a grounded base transistor. Correcting what I said earlier the grounded base configuration has low input impedance and is good at isolating its input from its output.

Perhaps Morse code should be brought back because it works with such narrow bandwidth carriers that can be filtered out of the worst of globally increasing noise levels.

It would be interesting to plot the signal strength of a far away VLF station for a few days to see how its signal strength varied with time from night time to daytime. Also how its affected by weather.
 
#160
I remember the 1st BJT radios because it was en vogue to design and build one. Like you said, with a low impedance tap or winding for the mixer. I'll never forget the disappointment, comparing it with my '7360 mixer' tube set 😉The low Z turn or tap plus the BJT's input C and the <1 coupling coefficient between windings on the ferrite, caused such receivers to be sensitive enough to produce an orchestra of "birdies" due to (unwanted) SW reception. So it was necessary to add an RF amp to "cure" that.
While CW has its merits, I disliked it enough not to learn it. RTTY still is a good contender for narrow band communications and with a proper algorithm produces so few errors that it's in the class of "good enough" for most purposes. So I treated FSK as a set of ASK signals in a way that if one of them went to zero due to fading or noise, copy remained 100%. The old UARTs used mid-bit sampling whereas I used 16 samples per bit whereby <8 "high" regarded "0", >8 "1" and =8 was interpreted depending on "history". That in turn was enough to program a sequence of characters not used in any language and when producing such a string, it was deleted. So even when operating for a long time without signal, I never saw a screen of rubbish (contrary to modern "all in software" ARQ and FEC decoders).
One very useful, never applied propagation mode for broadcasting is NVIS. It regards sending "straight up" and in the proper frequency range the reflection is strong. I never heard of it till receiving a Costa Rican station at 2860 KHz. Good copy with little fading. Turned out to be the 2nd harmonic of radio San Carlos.
This might be of interest on VLF: https://www.mdpi.com/2072-4292/15/4/1018
NVIS: https://en.wikipedia.org/wiki/Near_vertical_incidence_skywave
 
You must log in or register to reply here.
Top Bottom